On Sat, 25 Aug 2012 13:10:05 -0700
Diego Elio Pettenò wrote:
> On 25/08/2012 13:09, Michał Górny wrote:
> > That's not PMS-compliant, sir.
>
> But it's terribly useful. (I have expressed before my contempt for
> PMS, I'm sure.)
Didn't you get upset when it was Mike who was ignoring everyone else
On 25/08/2012 13:09, Michał Górny wrote:
> That's not PMS-compliant, sir.
But it's terribly useful. (I have expressed before my contempt for PMS,
I'm sure.)
I'm not expecting the ebuild to die, just to send a warning so that the
developer knows when they fail to meet non-obvious requirements.
--
On Sat, 25 Aug 2012 12:58:19 -0700
Diego Elio Pettenò wrote:
> On 25/08/2012 12:32, Michał Górny wrote:
> > + local
> > pkgconf=$(tc-getPKG_CONFIG)
>
> I'd say you should add a warning before using this if the virtual is
> not in DEPEND. And don't assume transit
On 25/08/2012 12:32, Michał Górny wrote:
> + local pkgconf=$(tc-getPKG_CONFIG)
I'd say you should add a warning before using this if the virtual is not
in DEPEND. And don't assume transitive dependencies (they are never nice).
You can take a peek to ruby-ng_rsp
Right now, the .pc file parsing takes place each time
prune_libtool_files() is called, even when there are no .la files
installed. After this commit, the .pc files will be parsed only when
it is necessary, i.e. no other criteria allow removal of a particular
.la file.
---
gx86/eclass/eutils.eclass